
\documentclass{article}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%TCIDATA{OutputFilter=LATEX.DLL}
%TCIDATA{Version=5.50.0.2960}
%TCIDATA{<META NAME="SaveForMode" CONTENT="1">}
%TCIDATA{BibliographyScheme=Manual}
%TCIDATA{Created=Friday, January 13, 2017 09:01:11}
%TCIDATA{LastRevised=Tuesday, January 17, 2017 12:43:27}
%TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
%TCIDATA{<META NAME="DocumentShell" CONTENT="Scientific Notebook\Blank Document">}
%TCIDATA{CSTFile=Math with theorems suppressed.cst}
%TCIDATA{PageSetup=72,72,72,72,0}
%TCIDATA{AllPages=
%F=36,\PARA{038<p type="texpara" tag="Body Text" >\hfill \thepage}
%}


\newtheorem{theorem}{Theorem}
\newtheorem{acknowledgement}[theorem]{Acknowledgement}
\newtheorem{algorithm}[theorem]{Algorithm}
\newtheorem{axiom}[theorem]{Axiom}
\newtheorem{case}[theorem]{Case}
\newtheorem{claim}[theorem]{Claim}
\newtheorem{conclusion}[theorem]{Conclusion}
\newtheorem{condition}[theorem]{Condition}
\newtheorem{conjecture}[theorem]{Conjecture}
\newtheorem{corollary}[theorem]{Corollary}
\newtheorem{criterion}[theorem]{Criterion}
\newtheorem{definition}[theorem]{Definition}
\newtheorem{example}[theorem]{Example}
\newtheorem{exercise}[theorem]{Exercise}
\newtheorem{lemma}[theorem]{Lemma}
\newtheorem{notation}[theorem]{Notation}
\newtheorem{problem}[theorem]{Problem}
\newtheorem{proposition}[theorem]{Proposition}
\newtheorem{remark}[theorem]{Remark}
\newtheorem{solution}[theorem]{Solution}
\newtheorem{summary}[theorem]{Summary}
\newenvironment{proof}[1][Proof]{\noindent\textbf{#1.} }{\ \rule{0.5em}{0.5em}}
\input{tcilatex}

\begin{document}


\begin{flushleft}
This file documents the data files and programs needed to reproduce the
results in Croushore and van Norden "Fiscal Forecasts at the FOMC: Evidence
from the Greenbooks", published in \textit{The Review of Economics and
Statistics.} 
\end{flushleft}

\section{Summary}

To replicate the results reported in the paper, place all the files in the
working directory. Then

\begin{enumerate}
\item Run \texttt{GB\_raw\_data.gau }and\texttt{\ Actuals.gau }in GAUSS

\item Run \texttt{Create\_GBFE.RPF} in WinRATS.

\item Run the WinRATS program that corresponds to the results that you wish
to reproduce

\begin{itemize}
\item \texttt{Table\_2.RPF}

\item \texttt{Table\_3.RPF}

\item \texttt{Table\_4.RPF}

\item \texttt{Table\_5.RPF }(also produces the results from Section II of
the Appendix.)

\item \texttt{Table\_6.RPF }(also produces Figures 4 \& 5 as well as the
results in Section III\ of the Appendix.)
\end{itemize}
\end{enumerate}

\section{Software}

Calculations were done using GAUSS 15.0.7 build 3835, WinRATS\ Standard
9.00f, RATSData 9.00, and Microsoft Excel 2013 running under Microsoft
Windows~7 Entreprise Version 6.1.7601 Service Pack 1 Build 7601.

\section{Data series}

We work primarily with six series: SURPLUS, EXPEND, RECEIPTS, HEB, HEB6 and
UNEMP as explained in Table 1 of the paper. Section I of the Appendix to the
paper (particularly in subsections A and F) provides extensive details on
the sources, definitions and preparation of our series. 

\section{Programs and Data files}

Data were taken from multiple original sources and were processed in
multiple steps. 

\begin{enumerate}
\item Hand-collected Greenbook data from page scans of original Greenbooks
(described in Section I.A of the Appendix to the paper) were stored in 
\texttt{GB\_raw\_data.xlsx}. These were then processed by the GAUSS program 
\texttt{GB\_raw\_data.gau} to produce series of forecasts at fixed forecast
horizons for the first and last FOMC\ meeting of the month. Along the way,
some series were spliced and series in nominal values were scaled by GNP or
GDP values for the same Greenbook vintage and calendar date. The last
Greenbook-reported value for each series in each quarter was also recorded.
Running the GAUSS program six times (once for each series) produces six
files containing formatted Greenbook forecasts: \texttt{GB\_Receipts.xlsx,
GB\_Expend.xlsx, GB\_Surplus.xlsx, GB\_HEB.xlsx, GB\_HEB6.xlsx }and\texttt{\
GB\_Unemp.xlsx. }The program also produces the graph shown in \textbf{Figure
3}, as well as comparable graphs for the other series.

\item Vintage published statistics from ALFRED were stored in \texttt{%
ALFRED\_series.xls} (details on the series may be found in the first sheet
of the file.) The program \texttt{Actuals.gau} does the necessary
transformations (scaling by GNP or GDP, conversion from monthly to
quarterly) for comparison to the Greenbook forecasts and selects vintages
appropriate for forecast evaluation (such as the first release and the
pre-benchmark release) as described in Section I.D of the Data Appendix.
Results were then stored in the file \texttt{Actuals.xlsx}$.$

\item The files produced from the two preceding steps were matched up in the
RATS program \texttt{Create\_GBFE.RPF} to produce series of forecast errors
for each forecast horizon and each outcome measure. Missing values were
inserted when benchmark changes intervened between the date of the forecast
and the date at which the official estimate was published. (The list of
benchmark revision dates may be found in Table 3 of the Appendix.) Results
were then stored in the files \texttt{GBFE\_SURPLUS.rat, GBFE\_RECEIPTS.rat,
GBFE\_EXPEND.rat, GBFE\_UNEMP.rat, GBFE\_HEB.rat, AND GBFE\_HEB6.rat.} 

\item The results shown in \textbf{Table 2} were produced by the WinRATS
program \texttt{Table\_2.RPF} using the data files produced in step 3,
above. 

\item The results shown in \textbf{Table 3} were produced by the WinRATS
program \texttt{Table\_3.RPF} and recorded in the file \texttt{Table\_3.xls}
using the data files produced in step 3. The values shown in the paper
appear in column F of the spreadsheet under the heading \textit{Fraction%
\TEXTsymbol{>}0} and p-values for the null hypothesis that this Fraction=0.5
are shown in column G under the heading \textit{p-Value}.

\item Results in \textbf{Table 4 }rely on hand-collected CBO forecasts
contained in the file \texttt{CBOHorseRacePreBench2016.RAT} as well as
annually aggregated Greenbook forecasts from the file \texttt{GB forecasts
to compare with CBO 2016 Aug alt.xlsx}. The latter were prepared by hand,
selecting the Greenbooks from the meeting date closest to the publication of
the CBO estimates and temporally aggregating quarterly estimates over the
current calendar year as well as the following calendar year. Calculations
are performed in the WinRATS program \texttt{Table\_4.RPF}. The RMSFEs
reported in the first two rows of Table 4 correspond to the values labelled
as RMSE(A) and RMSE(B) in the first line of the results reported for each
pair of forecasts. The p-values reported in the remaining rows of the Table
are taken from the last column of the results.

\item Results in \textbf{Table 5 }rely on additional series of Greenbook
forecast errors contained in the files \texttt{GBFE-PCPI.RAT}, \texttt{%
GBFE-PGDP.RAT}, \texttt{GBFE-RGDP.RAT} and \texttt{GBFE-YGAP.RAT}. Details
on the data sources used in the construction of these series may be found in
Section II of the Appendix. The results shown in Table 5 of the paper are
produced by the WinRATS program Table\_5.RPF together with further results
presented in Section II of the Appendix. Calculations are summarized in a
series of report windows, each of which follows the same format as the
tables from the paper and the Appendix.

\begin{itemize}
\item Column 1 indicates the forecast horizon, from shortest (0L) to longest
(4F)

\item Columns 2-4 show the results for the full sample

\item Columns 5-7 show the results for the pre-1991 sample

\item Columns 8-10 show the results for the post-1990 sample
\end{itemize}

\item Results in \textbf{Table 6} are produced by the WinRATS program 
\texttt{Table\_6.RPF} and the data file \texttt{Table\_6\_data.xlsx}. The
data file contains the data from Coibion et al. (2012) (as detailed in the
Appendix, Section III) as well as data from the files \texttt{%
GB\_Surplus.xlsx }and\texttt{\ GB\_HEB.xlsx}. The program collects results
in a number of report windows; report window \#1 contains the results
reported in \textbf{Table 6}, while report window \#2 contains the results
reported in Section III of the Appendix. It also produces the graphs shown
in \textbf{Figures 4 and 5 }(as well as comparable graphs for the other
models shown in Table 6.) 
\end{enumerate}

\section{Other Files}

The following files contain procedures used in some of the programs
mentioned above and must be available in the working directory when the
programs are run.

\begin{description}
\item[DCSvN2016.gcg] data handling utilities for GAUSS

\item[SignTests.src] Tests for non-zero medians.

\item[DieboldMariano.src] Diebold-Mariano and Harvey, et al. tests for
forecast performance and forecast encompassing.
\end{description}

\end{document}
